Senior AI judgement, on your side of the table.

Where are you right now?

Pick the sentence that sounds most like your team. We'll show you the engagement that fits — and why.

Explore: We know AI matters — but where do we start?

Recommended: Strategy Brief. Start with a Strategy Brief: we map 3–5 ranked AI opportunities against your data + stack and return a 90-day go/no-go roadmap.

Architect: We have a use case. We need the right architecture.

Recommended: Architectural Campaign. Run an Architectural Campaign: data-to-model selection, RAG/agent design, working prototype, and a production readiness plan.

Ship: We're mid-build. We need senior execution help.

Recommended: Architectural Campaign. Architectural Campaign with a sprint overlap — we plug into your team, ship the hard parts, and hand off a proven playbook.

Scale: We're live. We need optimization + ops.

Recommended: Embedded Ops. Go with Embedded Ops: a weekly working session, on-demand architecture reviews, continuous evals, and cost/latency tuning.

Three ways to work together. Priced up front.

Fixed fees or capped time-and-materials. No utilisation theatre, no surprise invoices, no lock-in.

Strategy Brief

Sprint · direction

1–2 weeks · from €5K fixed fee · NDA · no lock-in

Best for: For leaders who need a confident first move before committing budget.

• AI readiness assessment of your current stack + data
• 3–5 ranked opportunities with ROI + risk score
• Architecture options (RAG / agents / fine-tune) with trade-offs
• 90-day executable roadmap
• Clear go/no-go recommendation + hand-off artefacts

Architectural Campaign

Multi-week · build

4–8 weeks · from €25K fixed-scope or capped T&M

Best for: For teams with a chosen use case who need it designed right the first time — and shipped.

• Reference architecture (diagrams + decision log + ADRs)
• Data → model selection with evaluations against your tasks
• Prototype or production MVP built on your stack
• Evals, guardrails, observability + cost model
• Production readiness plan + team handover playbook

Embedded Ops

Retainer · senior copilot

3+ months · from €15K/mo monthly retainer · 30-day notice

Best for: For post-launch AI operators who need a senior force-multiplier embedded in the team.

• Weekly working sessions + async Slack/Linear presence
• On-demand architecture reviews + second opinions
• Continuous eval pipeline + regression catches
• Cost + latency + quality tuning
• Team enablement: playbooks, reviews, pairing sessions

How an engagement runs.

Understand before recommending. Compare options with numbers. Ship on your stack. Keep it sharp after launch.

P1 · Discover

Understand before recommending · 3–5 days

We read your codebase, map your data, meet the people who run it, and audit the stack. Nothing gets designed in a vacuum.

P2 · Architect

Option space, not opinions · 1–2 weeks

We draft 2–3 architecture options with RAG vs agents vs fine-tune trade-offs, run micro-benchmarks, and choose by numbers.

P3 · Deploy

Shipping is the deliverable · 2–6 weeks

We build the hard parts end-to-end with your team, wire up evals + guardrails, and bring it to production on your stack.

P4 · Operate

Keep it sharp in the wild · Ongoing

Weekly evals against regression suites, cost/latency tuning, model swaps when better ones ship, and playbooks for your team.

Results, not decks.

Three engagements, told the way a CFO would want to read them.

FinTech support concierge

FinTech

Problem: 80% of L1 tickets were repeat questions. CSAT falling.

Solution: Private RAG over policies + ticket history · tiered model routing · guardrails + escalation.

Outcome: −47% support cost · +18pt CSAT · 92% first-contact resolution

E-commerce Gen-AI pages

D2C retail

Problem: Generic PDPs eroding conversion vs niche competitors.

Solution: Generative personalised product pages · cached variants · A/B at edge.

Outcome: +3.2× conversion on top 500 SKUs · 73ms p95 latency · €1.4M GMV in 90 days

Clinical Q&A for staff

Healthcare

Problem: Clinicians wasting hours searching protocols + drug charts.

Solution: Hybrid RAG on curated corpus · attribution citations · offline eval harness.

Outcome: 92% factual accuracy · 400 staff onboarded in 8 weeks · 35min/day saved

Founder-led. Senior on every call.

The person who scopes your brief designs the architecture and ships the hard parts. No bait-and-switch, no rotating account managers.

Specialties

• AI Strategy & Roadmaps — Prioritisation, ROI, sequencing
• RAG / Agent Architecture — Retrieval, tools, memory, eval
• Model Selection + Evals — Objective benchmarks, not vibes
• Voice + Multi-modal — Latency, barge-in, handoff
• Production Deployment — GCP / AWS / Vertex / Bedrock
• Cost + Latency Tuning — Caching, routing, quantisation
• Safety, Privacy, Governance — PII, audits, SOC2 paths
• Team Enablement — Playbooks, reviews, workshops

Model-agnostic stack

Claude, GPT, Gemini, Llama · Mistral, LangGraph · DSPy, Vertex AI · Bedrock, Pinecone · pgvector, Next.js · FastAPI, GCP · AWS · Cloudflare, Firebase · Supabase

How we operate.

Delivery > decks

Every engagement ends in production or an implementation-ready blueprint. No 40-slide PDFs that die in a drive.

Specificity wins

One great use case beats ten mediocre ones. We kill the fluff and over-index on the highest-leverage opportunity.

Model-agnostic

We follow the evals, not the vendor. Your stack gets what performs today — and can swap what performs tomorrow.

Senior-only

Founders and staff engineers on every engagement. No junior handoff, no rotating account managers.

Transparent pricing

Fixed fees or capped T&M, priced per scope. No surprise invoices, no utilisation theatre.

Your stack, your repo

We build on what you already run. We leave when you tell us to — with playbooks that survive our departure.

Before you ask.

How much does a Strategy Brief cost?

Brief engagements start at €5K and conclude with an actionable go/no-go you can execute with any partner — no lock-in.

Do you build it, or just design it?

Both. We architect, prototype, and ship the hard parts. Production hand-off works with your team, or we stay embedded via retainer.

Which models do you work with?

Claude 4.6/4.7, GPT-5 / 4.1, Gemini 2.5 Pro/Flash, Llama 3.x, Mistral, Kimi, and fine-tuned open-source. Model-agnostic — the stack fits the use case.

Is this enterprise-ready?

Yes. SOC2 paths, PII redaction, VPC endpoints, private RAG, audit logs. Most engagements close with a compliance checklist signed off.

How fast can we start?

Strategy Briefs typically kick off within 5 business days of sign-off. Architectural Campaigns take ~10 days to schedule. Retainers require a 30-day onboarding.

Who is on the other side?

Founder-led. Every engagement is senior — you get the strategist who scopes it and the engineer who builds it, no bait-and-switch.

Do you sign NDAs?

Default. We send a mutual NDA with the first call invite. Your repo, data samples, and internal docs never leave your perimeter without it.

Can you work with our existing agency / dev shop?

Yes. A common pattern is us as senior architects; your agency as hands-on builders. We document clean handoffs so nobody is blocked.

Bring the hard question. Leave with a direction.

A strategy call is 30 minutes with the founder, with a mutual NDA by default.